Volume 89, Issue 2, Pages 297-308 (April 1997) MARK, a Novel Family of Protein Kinases That Phosphorylate Microtubule-Associated Proteins and Trigger Microtubule Disruption  Gerard Drewes, Andreas Ebneth, Ute Preuss, Eva-Maria Mandelkow, Eckhard Mandelkow  Cell  Volume 89, Issue 2, Pages 297-308 (April 1997) DOI: 10.1016/S0092-8674(00)80208-1

Figure 1 Mammalian Kinases MARK1 and MARK2 and Their Relatives from Nematode and Yeast (A) Alignment of the predicted amino acid sequences of MARK1 and MARK2 with PAR-1 (C. elegans; Guo and Kemphues 1995) and kin1+ (S. pombe; Levin and Bishop 1990) using CLUSTALW (Thompson et al. 1994). Identical residues are shaded in black, similar residues in grey. The activating phosphorylation sites in subdomain VIII are indicated by dots. Dashed lines above the sequence indicate porcine MARK peptides (Table 2). The catalytic domain is subdivided into subdomains I–XI. Within the tail domain, two predicted amphipathic helices are indicated by “a,” N713–L723 and I777–L793. (B) Diagram of the proposed domain structure of MARK kinases (see text). (C) Dendrogram of an alignment of the kinases of the SNF1/AMPK subfamily. Cell 1997 89, 297-308DOI: (10.1016/S0092-8674(00)80208-1)

Figure 1 Mammalian Kinases MARK1 and MARK2 and Their Relatives from Nematode and Yeast (A) Alignment of the predicted amino acid sequences of MARK1 and MARK2 with PAR-1 (C. elegans; Guo and Kemphues 1995) and kin1+ (S. pombe; Levin and Bishop 1990) using CLUSTALW (Thompson et al. 1994). Identical residues are shaded in black, similar residues in grey. The activating phosphorylation sites in subdomain VIII are indicated by dots. Dashed lines above the sequence indicate porcine MARK peptides (Table 2). The catalytic domain is subdivided into subdomains I–XI. Within the tail domain, two predicted amphipathic helices are indicated by “a,” N713–L723 and I777–L793. (B) Diagram of the proposed domain structure of MARK kinases (see text). (C) Dendrogram of an alignment of the kinases of the SNF1/AMPK subfamily. Cell 1997 89, 297-308DOI: (10.1016/S0092-8674(00)80208-1)

Figure 2 MARK Expression Is Pronounced in Brain, Spleen, Kidney and Muscle Northern blot analysis of 2 μg of mRNA from rat and human tissues. Size markers in kilobases are indicated. The upper blots show hybridization with MARK1-specific oligonucleotides. Expression is pronounced in brain and spleen, but also in skeletal muscle and kidney. The 6.5 kb band in muscle possibly represents an unrelated mRNA, since it does not hybridize to a cDNA probe for MARK (data not shown). The lower blots show hybridization with MARK2-specific probes. Expression is prominent in brain, spleen, muscle, and kidney. In humans, MARK-related mRNAs are prominent in heart and muscle. Relatively high levels of a MARK2-related mRNA are found in fetal tissues. Cell 1997 89, 297-308DOI: (10.1016/S0092-8674(00)80208-1)

Figure 3 MARK Is Phosphorylated on Ser, Thr, and Tyr Residues, and its Activity Depends on Phosphorylation of Thr-215 and Ser-219 (A) Silver-stained gel showing porcine MARK (lane 1), the same sample after a 20 min incubation with 10 μg/ml PP2A (lane 2), and a control where PP2A was inhibited by microcystin-LR (lane 3). Note the mobility shift concomitant with dephosphorylation. The additional higher mobility bands at 90 kDa are reproducibly formed and could result from an increased proteolytic susceptibility of the dephosphorylated protein. Lanes 4–7 show rat MARK1 and MARK2 translated in reticulocyte lysates, detected by blotting with MARK C-terminal antiserum. Lane 4, MARK1; lane 5, MARK1 dephosphorylated; lane 6, MARK2; lane 7, MARK2 dephosphorylated. Lanes 8–10 show a phosphotyrosine immunoblot. Brain MARK is tyrosine phosphorylated (lane 8), which is sensitive to PTP1B treatment (lane 9) but not to PP2A (lane 10). (B) Brain MARK activity as dependent on autophosphorylation or dephosphorylation. Purified MARK was incubated with either 100 μM ATP (closed circles), PTP1B (10 μg/ml, open squares), or PP2A (10 μg/ml, open circles). After the indicated time, aliquots were tested for their ability to phosphorylate peptide TR1. Tyr dephosphorylation does not cause a significant decrease in activity, whereas the dephosphorylation of Ser/Thr residues results in complete inactivation. (C) Alignment of the sequence of a phosphopeptide isolated from porcine brain MARK (Table 2) with the corresponding sequences from several kinases. The phosphorylated Thr-215 in MARK corresponds to the regulatory Thr residue in PKA, CDKs, ERKs, and MEKs. (D) Point mutations in subdomain VIII affect MARK activity. The activating phosphorylation sites and an essential Lys residue were mutated in MARK1 and MARK2 plasmids as indicated. From these, synthetic RNAs were translated in reticulocyte lysate, and the products immunoprecipitated with anti-HA and anti-C-terminal antibodies. A blot of the samples with the C-terminal antibody is shown. The dephosphorylated sample (lane 2) and the mutants (lanes 4–7 and 9–12) display increased electrophoretic mobility. Catalytic activity was assayed by autoradiography after incubation with [γ-32P]ATP and tau. Only the wild-type enzymes phosphorylate tau efficiently. (E) Specificity of recombinant MARK. Tryptic digests of the tau samples phosphorylated with MARK1 (lane 8 in [D]) or MARK2 (lane 3 in [D]) were mapped by thin-layer electrophoresis (horizontal) followed by ascending chromatography (vertical). The main spots correspond to the four KXGS motifs in the tau repeats (Drewes et al. 1995). For comparison, a map of tau phosphorylated with porcine brain MARK is shown. Cell 1997 89, 297-308DOI: (10.1016/S0092-8674(00)80208-1)

Figure 4 Overexpression of MARK in CHO Cells Causes Breakdown of the Microtubule Array Double immunofluorescence microscopy of HA-tagged MARKs transfected into CHO cells. After 24 hr, cells were fixed with methanol and stained with anti-HA (left panels) and anti-tubulin (YL1/2) antibodies (right panels). Scale bar = 5 μm. (A) MARK2-transfected cells tend to be smaller and more rounded in shape, and eventually die. (B) MARK2-transfected cells showing a dramatic loss of MTs. (C) Transfection with catalytically inactive MARK2T208A/S212A. In contrast to wild-type MARK2, no effect on the MT network is visible. (D) Cells transfected with MARK1. MT destruction is also observed (arrowheads) but is less pronounced compared to MARK2 overexpression. The staining pattern of the transfected MARK1 is clearly distinct from that of MARK2. (E) Transfection with inactive MARK1T215A/S219A. No effect on the MT network is visible. Cell 1997 89, 297-308DOI: (10.1016/S0092-8674(00)80208-1)

Figure 5 The Phenotypical Effects of MARK Can Be Blocked by Taxotere or Coexpression of MAP2c (A) Time course of the appearance of morphologically abnormal cells after transfection with MARK2 (open circles), MARK1 (squares) and the inactive MARK2T208A/S212A mutant (closed circles). Triangles show mock-tranfected cells for control. Cells were analyzed by immunofluorescence with anti-HA and anti-tubulin antibodies and regarded as “morphologically abnormal” if they were less than 30% of the average size of untransfected cells, with no distinct MT network visible. (B) Blocking of MARK2 effects by the microtubule-stabilizing drug taxotere. Cells were counted 18 hr after transfection, as in (A), in parallel experiments with and without 10 μM taxotere. (C) Partial blocking of MARK2 effects by cotransfection with MAP2c. MAP2cS319A/S350A, a mutant in which the two main MARK phosphorylation sites have been eliminated, was more effective in counteracting MARK2 effects than wild-type MAP2c. About 400 cells were counted in each experiment, and mean values of three independent experiments ± SD are shown. Cell 1997 89, 297-308DOI: (10.1016/S0092-8674(00)80208-1)

Figure 6 MARK Expression Leads to Hyperphosphorylation of Tau on KXGS Sites Tau was transiently transfected into CHO cells that stably express MARK2 under control of a doxycyclin-inducible promoter. Cells were fixed with formaldehyde and stained with anti-tau plus anti-tubulin (left panels) and a phospho-tau antibody specific for KXGS-type phosphorylation, 12E8 (right panels). Upper panels show cells in which MARK2 expression is induced, lower panels show the control without induction. The left panels show double staining with tau and tubulin antibodies. Note that, after induction of MARK, the cells that do not express tau (arrowheads) show a weakened microtubule array, compared to the control where no MARK expression had been induced (lower left panel). The bright cells express tau (they are double labeled with tau and tubulin). The microtubule network of the tau-expressing cells remains more intact, presumably because the overexpression of MAPs stabilizes microtubules. The right panels show the phosphorylation of KXGS motifs on tau, as detected by the phosphorylation-dependent antibody 12E8. Phosphorylation of tau at KXGS motifs is observed when MARK expression is induced (upper right panel) but not in the control without doxycyclin induction (lower right panel). Cell 1997 89, 297-308DOI: (10.1016/S0092-8674(00)80208-1)